ESM-Snow model intercomparison

Transcription

1 Grand Challenge Cryosphere in a Changing Climate ESM-Snow model intercomparison C. Derksen, G. Krinner, R. Essery, M. Flanner, S. Hagemann, H. Rott

2 Motivation Rapid snow extent changes in NH Climate change indicator & climate model verifica:on parameter Snow poten:ally important for climate predic:on on seasonal scale Snow feedbacks are key for longer :me scales RS: Recent progress (SnowPEX) Possibility to build on progress in dedicated snow modelling (SnowMIP) The current strengths and weaknesses of snow models used in ESMs must be assessed in order to provide guidelines for their improvement Bring together site and large scale modeling community, site and large-scale observa?ons

3 Issues with snow in ESMs Representation of vertical snow variability and fluxes: Number of layers, vertical discretization, Snow fraction parameterisation: depends on the season and vegetation Albedo parameterisation: prognostic vs. diagnostic, black carbon Snow-vegetation interaction: including multi-energy balance? Snow density and its impact on heat conductivity Blowing snow and associated impact on sublimation Heat conductivity: major impact on underlying soil (Gouttevin et al., 2012)

4 Consequences of these issues Example: Permafrost extent Snow extent not that bad, but underlying soil temperatures vary widely. Reason for misfits: soil + snow physics (Koven et al., 2013) (Flato et al., 2013 [IPCC AR5 Ch. 9])

5 Example: Simulated snow feedback "The spread in snow albedo feedback is very similar to that found in CMIP3 models, and it accounts for much of the spread in the 21 st century warming of Northern Hemisphere land masses in the CMIP5 ensemble, especially in spring and early summer. Qu and Hall (2013), doi: /s

6 What we learned from previous snow model intercomparisons Site-level, with dedicated models Models capture broad features of snow accumula:on and abla:on Broad spread between models, par:cularly in warmer condi:ons There is no «best» model Model performance is not clearly related to model complexity Driving and evalua:on data errors are hard to separate from model errors Interpreta:on of results complicated by different interpreta:on of instruc:ons and different degrees of calibra:on But much of the spread can be reproduced in mul:-physics ensembles, with more physically-based parameteriza:ons performing berer

7 A word on (global) snow climatologies There is considerable inter-dataset spread in Northern Hemisphere snow mass and snow cover extent derived from available terrestrial snow products Efforts are underway (through ESA SnowPEx) to derive an op:mal ensemble of observed products in order to provide an observa:onal founda:on for CMIP6 land MIPs (i.e. LS3MIP) Mudryk et al (2015) J. Climate

8 Con:nental and Hemispheric Satellite Snow Extent Products Name Product type Pixel Spacing Frequency Period Main Sensor Organisa?on NOAA IMS Binary 4 km daily NOAA IMS Binary 24 km daily GlobSnow MOD10 Frac:onal FSC Frac:onal FSC 1 km daily - monthly OPT, PMW, OPT, PMW ATSR2 AATSR 0.5 km Daily MODIS NOAA (Helfrich et al.) NOAA SYKE, ENVEO et al. ESA support NASA / NSIDC (Hall et al.) AVHRR Pathfinder Binary 5 km daily AVHRR CCRS (Zhao, et al) CryoLand Frac:onal (Europe) 0.5 km daily MODIS ENVEO / SYKE et al.- EC -FP7

9 Main questions How bad are current snow models (specific ones & those used in climate modeling)? What processes do they have to represent in a climate model? How strong are, and will be, snow-related climate feedbacks (real & model world)? Axes Evalua:on of snow models against new, diverse and longer site observa:ons: Extend SnowMIP 1&2 Evaluate snow in CMIP6 models (coupled and LMIP) link to LS3MIP Quan:fy snow feedbacks in CMIP6 simula:ons link to LS3MIP

10 Site simulations Tier 1: Reference site simulations Tier 2: Shallow soil Large-scale forcing Fixed albedo High thermal conductivity

11 Global simulations Tier 1: Forced (uncoupled) global simulations: Fixed snow albedo Prescribed observed snow water equivalent Coupled: LS3MIP land forcing simulations restricted to snow: total snow effect snow effect on SW radiation only

12 Snow radiative effect Instantaneous perturbation to Earth s solar energy budget induced by the presence of surface cryospheric components Diagnosed through parallel radiation calculations of surface albedo and solar energy fluxes with and without the presence of snow Analogous to cloud radiative effect Northern Hemisphere (NH) CrRE derived from a variety of remote sensing data (MODIS, AVHRR, AMSR- E) Change in NH CrRE during : ( ) W m 2, half of which was caused by reduced terrestrial snow Include snow radiative effect calculations in ESM-SnowMIP/LS3MIP global simulations (Flanner et al., 2011)

13 Planning Linked to LS3MIP: global ESM-SnowMIP simulations complementary Finalize site forcing data (now) Kick-off workshop Dec 10, Fort Mason, San Francisco (pre-agu) Start site simulations Global off-line and coupled simulations: After CMIP6 (LS3MIP) Long term: snow on sea ice snow on ice sheets.